Skip to main content
. 2011 Jul 7;18(5):574–579. doi: 10.1136/amiajnl-2011-000302

Table 3.

Scores for challenge test, training set and a baseline model for relation classification

Entity type Training Testing Recall (test), recall (train), (baseline) Precision (test), precision (train), (baseline) F-score (test), F-score (train), (baseline)
PIP 1239 1986

  • 62.74%

  • 64.32%

  • (62.95%)

  • 67.68%

  • 72.95%

  • (69.09%)

  • 65.12%

  • 67.91%

  • (65.88%)
TrWP 56 143

  • 2.80%

  • 3.57%

  • (3.57%)

  • 80%

  • 100%

  • (100%)

  • 5.41%

  • 6.90%

  • (6.90%)
TrAP 1422 2487

  • 72.46%

  • 77.92%

  • (77.57%)

  • 69.90%

  • 68.48%

  • (63.68%)

  • 71.15%

  • 72.89%

  • (69.94%)
TrNAP 106 191

  • 13.09%

  • 26.42%

  • (25.47%)

  • 55.56%

  • 70%

  • (71.05%)

  • 21.19%

  • 38.36%

  • (37.50%)
TrCP 296 444

  • 47.97%

  • 44.93%

  • (47.64%)

  • 49.53%

  • 63.64%

  • (62.95%)

  • 48.74%

  • 52.67%

  • (54.23%)
TrIP 107 198

  • 15.66%

  • 23.36%

  • (25.23%)

  • 86.11%

  • 69.44%

  • (64.29%)

  • 26.50%

  • 34.97%

  • (36.24%)
TeCP 303 588

  • 43.03%

  • 47.85%

  • (44.88%)

  • 61.41%

  • 77.13%

  • (74.32%)

  • 50.60%

  • 59.06%

  • (55.97%)
TeRP 1733 3033

  • 84.04%

  • 86.96%

  • (87.31%)

  • 84.04%

  • 82.39%

  • (79.93%)

  • 84.04%

  • 84.62%

  • (83.45%)
Overall 5262 9070

  • 67.51%

  • 70.90%

  • (70.87%)

  • 73.07%

  • 74.44%

  • (71.12%)

  • 70.18%

  • 72.63%

  • (70.99%)

PIP, problem indicates problem; TeCP, test conducted to investigate problem; TeRP, test reveals problem; TrAP, treatment is administered problem; TrCP, treatment causes problem; TrIP, treatment improves problem; TrNAP, treatment not administered because of problem; TrWP, treatment worsens problem.